Abstract. Advances in ambient environmental monitoring technologies are enabling concerned communities and citizens to collect data to better understand their local environment and potential exposures. These mobile, low-cost tools make it possible to collect data with increased temporal and spatial resolution, providing data on a large scale with unprecedented levels of detail. This type of data has the potential to empower people to make personal decisions about their exposure and support the development of local strategies for reducing pollution and improving health outcomes. However, calibration of these low-cost instruments has been a challenge. Often, a sensor package is calibrated via field calibration. This involves colocating the sensor package with a high-quality reference instrument for an extended period and then applying machine learning or other model fitting technique such as multiple linear regression to develop a calibration model for converting raw sensor signals to pollutant concentrations. Although this method helps to correct for the effects of ambient conditions (e.g., temperature) and cross sensitivities with nontarget pollutants, there is a growing body of evidence that calibration models can overfit to a given location or set of environmental conditions on account of the incidental correlation between pollutant levels and environmental conditions, including diurnal cycles. As a result, a sensor package trained at a field site may provide less reliable data when moved, or transferred, to a different location. This is a potential concern for applications seeking to perform monitoring away from regulatory monitoring sites, such as personal mobile monitoring or high-resolution monitoring of a neighborhood. We performed experiments confirming that transferability is indeed a problem and show that it can be improved by collecting data from multiple regulatory sites and building a calibration model that leverages data from a more diverse data set. We deployed three sensor packages to each of three sites with reference monitors (nine packages total) and then rotated the sensor packages through the sites over time. Two sites were in San Diego, CA, with a third outside of Bakersfield, CA, offering varying environmental conditions, general air quality composition, and pollutant concentrations. When compared to prior single-site calibration, the multisite approach exhibits better model transferability for a range of modeling approaches. Our experiments also reveal that random forest is especially prone to overfitting and confirm prior results that transfer is a significant source of both bias and standard error. Linear regression, on the other hand, although it exhibits relatively high error, does not degrade much in transfer. Bias dominated in our experiments, suggesting that transferability might be easily increased by detecting and correcting for bias. Also, given that many monitoring applications involve the deployment of many sensor packages based on the same sensing technology, there is an opportunity to leverage the availability of multiple sensors at multiple sites during calibration to lower the cost of training and better tolerate transfer. We contribute a new neural network architecture model termed split-NN that splits the model into two stages, in which the first stage corrects for sensor-to-sensor variation and the second stage uses the combined data of all the sensors to build a model for a single sensor package. The split-NN modeling approach outperforms multiple linear regression, traditional two- and four-layer neural networks, and random forest models. Depending on the training configuration, compared to random forest the split-NN method reduced error 0 %–11 % for NO2 and 6 %–13 % for O3.
more »
« less
Distributed Temperature Sensing for Oceanographic Applications
Abstract Distributed temperature sensing (DTS) uses Raman scatter from laser light pulsed through an optical fiber to observe temperature along a cable. Temperature resolution across broad scales (seconds to many months, and centimeters to kilometers) make DTS an attractive oceanographic tool. Although DTS is an established technology, oceanographic DTS observations are rare since significant deployment, calibration, and operational challenges exist in dynamic oceanographic environments. Here, results from an experiment designed to address likely oceanographic DTS configuration, calibration, and data processing challenges provide guidance for oceanographic DTS applications. Temperature error due to suboptimal calibration under difficult deployment conditions is quantified for several common scenarios. Alternative calibration, analysis, and deployment techniques that help mitigate this error and facilitate successful DTS application in dynamic ocean conditions are discussed.
more »
« less
- NSF-PAR ID:
- 10213700
- Date Published:
- Journal Name:
- Journal of Atmospheric and Oceanic Technology
- Volume:
- 37
- Issue:
- 11
- ISSN:
- 0739-0572
- Page Range / eLocation ID:
- 1987 to 1997
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Low-cost sensors enable finer-scale spatiotemporal measurements within the existing methane (CH 4 ) monitoring infrastructure and could help cities mitigate CH 4 emissions to meet their climate goals. While initial studies of low-cost CH 4 sensors have shown potential for effective CH 4 measurement at ambient concentrations, sensor deployment remains limited due to questions about interferences and calibration across environments and seasons. This study evaluates sensor performance across seasons with specific attention paid to the sensor's understudied carbon monoxide (CO) interferences and environmental dependencies through long-term ambient co-location in an urban environment. The sensor was first evaluated in a laboratory using chamber calibration and co-location experiments, and then in the field through two 8 week co-locations with a reference CH 4 instrument. In the laboratory, the sensor was sensitive to CH 4 concentrations below ambient background concentrations. Different sensor units responded similarly to changing CH 4 , CO, temperature, and humidity conditions but required individual calibrations to account for differences in sensor response factors. When deployed in-field, co-located with a reference instrument near Baltimore, MD, the sensor captured diurnal trends in hourly CH 4 concentration after corrections for temperature, absolute humidity, CO concentration, and hour of day. Variable performance was observed across seasons with the sensor performing well ( R 2 = 0.65; percent bias 3.12%; RMSE 0.10 ppm) in the winter validation period and less accurately ( R 2 = 0.12; percent bias 3.01%; RMSE 0.08 ppm) in the summer validation period where there was less dynamic range in CH 4 concentrations. The results highlight the utility of sensor deployment in more variable ambient CH 4 conditions and demonstrate the importance of accounting for temperature and humidity dependencies as well as co-located CO concentrations with low-cost CH 4 measurements. We show this can be addressed via Multiple Linear Regression (MLR) models accounting for key covariates to enable urban measurements in areas with CH 4 enhancement. Together with individualized calibration prior to deployment, the sensor shows promise for use in low-cost sensor networks and represents a valuable supplement to existing monitoring strategies to identify CH 4 hotspots.more » « less
-
Abstract. Exchanges between groundwater and surface water play a key role for ecosystem preservation, especially in headwater catchments where groundwater discharge into streams highly contributes to streamflow generation and maintenance. Despite several decades of research, investigating the spatial variability in groundwater discharge into streams still remains challenging mainly because groundwater/surface water interactions are controlled by multi-scale processes. In this context, we evaluated the potential of using FO-DTS (fibre optic distributed temperature sensing) technology to locate and quantify groundwater discharge at a high resolution. To do so, we propose to combine, for the first time, long-term passive DTS measurements and active DTS measurements by deploying FO cables in the streambed sediments of a first- and second-order stream in gaining conditions. The passive DTS experiment provided 8 months of monitoring of streambed temperature fluctuations along more than 530 m of cable, while the active DTS experiment, performed during a few days, allowed a detailed andaccurate investigation of groundwater discharge variability over a 60 m length heated section. Long-term passive DTS measurements turn out to bean efficient method to detect and locate groundwater discharge along several hundreds of metres. The continuous 8 months of monitoring allowed the highlighting of changes in the groundwater discharge dynamic in response to the hydrological dynamic of the headwater catchment. However, the quantification of fluxes with this approach remains limited given the high uncertainties on estimates, due to uncertainties on thermal properties and boundary conditions. On the contrary, active DTS measurements, which have seldom been performed in streambed sediments and never applied to quantify water fluxes, allow for the estimation of the spatial distribution of both thermal conductivities and the groundwater fluxes at high resolution all along the 60 m heated section of the FO cable. The method allows for the description of the variability in streambed properties at an unprecedented scale and reveals the variability in groundwater inflows at small scales. In the end, this study shows the potential and the interest of the complementary use of passive and active DTS experiments to quantify groundwater discharge at different spatial and temporal scales. Thus, results show that groundwater discharges are mainly concentrated in the upstream part of the watershed, where steepest slopes are observed, confirming the importance of the topography in the stream generation in headwater catchments. However, through the high spatial resolution of measurements, it was also possible to highlight the presence of local and highly contributive groundwater inflows, probably driven by local heterogeneities. The possibility to quantify groundwater discharge at a high spatial resolution through active DTS offers promising perspectives for the characterization of distributed responses times but also for studying biogeochemical hotspots and hot moments.more » « less
-
null (Ed.)Observing multiple size classes of organisms, along with oceanographic properties and water mass origins, can improve our understanding of the drivers of aggregations, yet acquiring these measurements remains a fundamental challenge in biological oceanography. By deploying multiple biological sampling systems, from conventional bottle and net sampling to in situ imaging and acoustics, we describe the spatial patterns of different size classes of marine organisms (several microns to ∼10 cm) in relation to local and regional (m to km) physical oceanographic conditions on the Delaware continental shelf. The imaging and acoustic systems deployed included (in ascending order of target organism size) an imaging flow cytometer (CytoSense), a digital holographic imaging system (HOLOCAM), an In Situ Ichthyoplankton Imaging System (ISIIS, 2 cameras with different pixel resolutions), and multi-frequency acoustics (SIMRAD, 18 and 38 kHz). Spatial patterns generated by the different systems showed size-dependent aggregations and differing connections to horizontal and vertical salinity and temperature gradients that would not have been detected with traditional station-based sampling (∼9-km resolution). A direct comparison of the two ISIIS cameras showed composition and spatial patchiness changes that depended on the organism size, morphology, and camera pixel resolution. Large zooplankton near the surface, primarily composed of appendicularians and gelatinous organisms, tended to be more abundant offshore near the shelf break. This region was also associated with high phytoplankton biomass and higher overall organism abundances in the ISIIS, acoustics, and targeted net sampling. In contrast, the inshore region was dominated by hard-bodied zooplankton and had relatively low acoustic backscatter. The nets showed a community dominated by copepods, but they also showed high relative abundances of soft-bodied organisms in the offshore region where these organisms were quantified by the ISIIS. The HOLOCAM detected dense patches of ciliates that were too small to be captured in the nets or ISIIS imagery. This near-simultaneous deployment of different systems enables the description of the spatial patterns of different organism size classes, their spatial relation to potential prey and predators, and their association with specific oceanographic conditions. These datasets can also be used to evaluate the efficacy of sampling techniques, ultimately aiding in the design of efficient, hypothesis-driven sampling programs that incorporate these complementary technologies.more » « less
-
null (Ed.)Recent waste rock pile designs have been proposed to incorporate a fine-grained layer to create a capillary barrier to prevent surface water from draining into the pile interior. This study analyses active fibre optic distributed temperature sensing (FO-DTS) as a tool to measure the effectiveness a capillary barrier system following an infiltration test. A laboratory waste rock column was built with anorthosite waste rock overlain by sand. Volumetric water content is calculated during heat cycles lasting 15 min powered at 15 W/m in the column. A new algorithm is employed to circumvent several requirements for soil specific calibration. The inferred moisture contents were verified by soil moisture probes located adjacent to the cable. The FO-DTS data indicate, at vertical resolutions up to 2 cm, that water is retained in the sand and does not drain into the anorthosite following the infiltration test. The coefficient of determination, R 2 , between the inferred and measured volumetric water content in the fine cover sand layer is 0.90, while the screened anorthosite maintained an R 2 of 0.94 with constant moisture content throughout the test. This study will ultimately help guide future waste rock storage design initiatives incorporating fibre optic sensors, leading to improved environmental mine waste management.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1175/JTECH-D-20-0066.1
